 So, this is the topic I was tasked to discuss with you and I would also like to thank everybody who came before me because we've discussed many of the elements that go into this synthesis in the days before this one, so this is what I would like to do. My research is really mainly on the West African monsoon, so that will be most of the work that I will discuss is based on my research, but also the research of Catherine who comes after me, who works with Yochanan and me at Le Monde and many others. Let's say the work on Eastern Equatorial Africa, we're talking about the decadal influences on these two rainfall systems. The work on Eastern Equatorial Africa is very much the work of a colleague of mine at IRI, Brad Lyon. But what I will try to do is try to bring some synthesis to how we look at these two rainfall systems, these two tropical climate systems in regards to decadal influences, influences on decadal timescale which come from the oceans. So the decadal timescale is very much dictated by the oceans and by variations on that timescale in the oceans and if I can summarize, I'll give you some of the conclusions already. It's mainly the Atlantic for the West African monsoon and the Pacific for Eastern Equatorial Africa. I would also like to bring some dynamics of tropical climates to the discussion. So try to go through some of the basic diagnostics of the variability that allow us also to some extent to bring some coherence across timescales from sub-seasonal to multidicadal and beyond. And I just put in gray here anthropogenic influence because the influence of the oceans on these climates can also be used to argue for or against anthropogenic influence. So a lot of the physics of this and the theory of this has to do with deep convection and vertical instability and what sets the instability that triggers deep convection in these tropical regions. And what is the role of the oceans? So what is the role of the oceans at the very most basic, so this is just annual mean precipitation over Africa and so the two regions are the regions that I will be discussing are the regions in the two boxes here. You can see based on climatology that we're talking about margins of climatological precipitation. So these are semi-arid regions and if you've been there, you know that my colleague from AgriMet likes to say, it really doesn't rain there for nine months of the year. So the moisture has to come from the ocean so at the very basic level we understand that the oceans have a very important role to play in the climate of these regions. Okay, so I'll go through some of the steps here and some of the motions in the case of West Africa in Monsoon and then try to repeat the same exercise for Eastern Equatorial Africa. So as has been pointed out already in the previous days, I mean this is one of the really, when we talk about decadal variability and predictability, the Sahel is really the most outstanding case of observed decadal variability in the 20th century instrumental record. And so what you see here in the red is a time series of the average of stations across the region from GHCN, anomalies, standardized anomalies in precipitation for this broad Sahelian region that was in that box that I showed on the map and so you see, I mean it's evident that there is multi-decadal variability and that there were decades of anomalously wet in the 50s and 60s followed by anomalously dry, persistent anomalously dry in the 70s and 80s and then perhaps some indication of a recovery in the most recent period. Now, there has been a long standing debate as to what caused these multi-decadal variations but I think now we can say with some certainty based on ensembles of AMIP simulations that these variations were caused by the oceans. So take an atmospheric model, force it with observed sea surface temperatures and you get something like the blue curve here. So knowledge of the observed global observed sea surface temperatures is sufficient to reproduce not just the decadal variations I would argue but also quite a bit of the inter-annual variations. So that's where you are. And then being that I work at IRI, this you can understand is the basis for seasonal prediction certainly perhaps also I guess if we could do decadal predictions then you can immediately see that there will be some value to those for this region. Okay, so how do we understand the role of the oceans? Yeah, that's the question one is prompted to ask here, okay, it's the global oceans but which oceans and how does this influence work? So I skipped a step here but if you look at, if you were to correlate the time series like the ones showed in the previous map to sea surface temperatures then you would see broad sensitivity of the significant correlations of the rainfall time series to sea surface temperatures in all the ocean basins. And here I broke it down by time scale if you wish into a long multi-decadal time scale and an inter-annual time scale and the moment I do that I can see that the inter-annual time scale is taken care of by ENSO. So here the blue means a negative correlation with Sahel rainfall. So positive anomalies where you see the blue mean drought in the Sahel. So El Nino typically causes drought in the Sahel most of the times that's the red bars on this plot. And if you're looking for an explanation for this multi-decadal variation then it's in the oceans around Africa. And so Indian Ocean Warming that Roxy just described is definitely one component and then the other component would be some kind of an inter-hemispheric gradient in the Atlantic. Okay so a little bit of how this could work specifically to well not just the Atlantic but let's take this initial figure which is the I guess the conventional wisdom about how rainfall varies in the Sahel says that if you have a gradient in sea surface temperatures of this nature so with a cool North Atlantic with respect to a warmer South Atlantic then this inter-hemispheric gradient drives the latitudinal location of the inter-tropical convergent zone and so when it's in this phase the ITCZ shifts southward and then you just extrapolate this to the continent and you get a dry Sahel. So I would like to offer a reinterpretation and a reinterpretation that moves us in the direction also trying to understand how climate change and warming may play into this and the interpretation is as follows it has two basic ingredients and that come from the theoretical work of all these people plus a few more perhaps I should have included for example also Adam Sobel or Lorenzo Polvani and to how so how does a tropical convection respond to these very broad-scale sea surface temperature variations. So one element is the global tropical oceans and you can think of El Nino as an analog so what happens if you warm the global tropical oceans as it happens during an El Nino event then what happens is that deep convection communicates that warming to the upper troposphere and so you affect the vertical stability by warming the upper troposphere you make environmental conditions more stable and so you you up the anti for convection to be triggered then in places that are remote from that deep tropical convection that has increased or strengthened. So then that would drive what we see as widespread drought during the development of an El Nino event and so the second ingredient and and you it would also so that would be for example the element of Indian Ocean warming affecting contributing to persistent drought in the Sahel. The second element is the moisture then how do you change that vertical stability how can you counter that increased stability that is communicated by deep convection through upper tropospheric warming and the idea here is that you counter it with the moisture that comes from the local ocean so in this case the Atlantic and so in this schematic of a cooler north to compare to a warmer south Atlantic then the cooler north Atlantic is not able to meet that up to NT that is driven by upper tropospheric warming because there's not enough moisture being driven into the monsoon and so you end up with a dryer Sahel in this configuration. Now if the climate were not warming then we wouldn't have to worry about this increasing trend in the warming so it's just that because we're in this transient warming that we have to consider the difference of the two. So when we consider the simple difference of these two oceanic influences which is in the then just simply captured in the difference between north subtropical north Atlantic sea surface temperatures and global tropical mean sea surface temperatures then we can explain the past variations so what do I have here I have these two elements here global tropical sea surface temperature anomaly on the x-axis and north Atlantic sea surface temperature anomaly on the y-axis and so you can see that my wet ears are in are when north Atlantic is warmer than global tropics so above this line that's full dots are positive anomalies and conversely the dry ears are below the line so north Atlantic not as warm or cooler than the global tropics and this on the right here is just how these ears are distributed in time so the blue dots are the wet epoch and the red dots are the dry epoch and now we're in this grey epoch where things are somewhat more variable here up on top so both north Atlantic and global tropics warming and now I just to show that how do we string along this convection that happens at very small spatial and temporal scales to sub seasonal characteristics of rainfall to then seasonal and then inter-annual multi-decadal and so on here I broke down rainfall into frequency on the left and intensity on the right and this is stations from the network of the National Meteorological Agency of Senegal and so same depiction here and what do we see now that the wet years are characterized by more rainy days compared to the dry years okay this is something that is well known so the wet years again the 50s and 60s had more rainy days during the rainy season than the dry years and now we're in this upper right corner so rainy days that are somewhat somewhere in between perhaps a little closer to the dry epoch than than the wet epoch in terms of number of rainy days and what is intriguing here is that the intensity actually what it really shows is more of more intense rains so more intense daily accumulation of rain that's in some ways compensates for the fewer rainy days so the recovery is really more of a recovery that is dictated by the by a slight increase in the intensity rather than in the number of days which is interesting I guess that's it follows expectation from climate change just generically speaking I know that Gavin is going to say something about that for sure okay so this it's how am I doing with time okay good so we know this very well so this is the time series of right the Sahelian rainfall so again the wet 50s and 60s the dry 70s and 80s and now this variable period that we're in so how do we explain this now coming coming to decadal variations and but I really want to remark the fact that it's it was really persistently wet in the 50s and 60s and sort of persistently dry and now we're in this really increasingly variable and epoch which again from the perspective of trying to apply the climate science really makes for an increased value I would say in a in a seasonal prediction system that can skillfully say something about how do you go from wet to dry year to year okay so I'm super imposing now these global multi-decadal oceanic influences the blue curve is an index of North Atlantic minus global mean so something very much like an AMO or an AMV index so North Atlantic sea surface temperatures detrended by subtracting the global mean temperature and so we see it has these nice oscillations that we've seen already the the green curve I should say is a smoothed version of Sahel rainfall and then the red curve here is broadly speaking equatorial Indian Ocean index with the sign reversed that's because the association is warming and with drying so I think you already get the sense that I can explain then the fact that it was persistently wet because both influences were in the direction of wet in the 50s and 60s whoops I am in trouble persistently dry in the 70s and 80s because both influences were in that direction and now we're in this phase because there's opposition of the two influences okay so that's the Sahel now let me try and wrap this up really quickly for Eastern equatorial Africa but there is an interesting the interesting decadal to longer term change and question here is that Eastern equatorial Africa has been persistently dry for about a decade or so but long-term projections say that it should get wetter so how do we reconcile these these two elements and I'm not going to try to do that but I'll just try and give an explanation for what's for what's going on at present so first an inter-annual timescale we understand so here the what what are the local and remote basins in relation to Eastern equatorial African rainfall so this is so in observations we see this relationship we're talking here this is the short range October November December there's a relatively clear association with the ENSO such that El Nino drives wet anomalies but it's mediated by an Indian Ocean warming so you get wet anomalies in Eastern equatorial Africa during El Nino if the Indian Ocean warms and so this is in observations in the top you get the same in Goga which was already talked about a global ocean global atmosphere but you don't get the same if you only have tropical Pacific warming with only tropical Pacific warming the atmospheric bridge would say dry Eastern equatorial Africa so you need this moisture you could cast this as a you need this added moisture that comes from the Indian Ocean that's supplied that is supplied into Eastern Africa to get wet maybe I'll skip this but to say that there's also but this breakdown of the seasonal rainfall into frequency and intensity also works to some extent here only now we're in the say let's say in the mature phase of ENSO so during El Nino's we get wetter conditions in Eastern equatorial Africa because we have more rainy days this is the stronger signal and but to some extent also more intense but it's not as strong a signature okay now on decadal time scale this is what has happened well it's a broader I mean you can see the signature in Eastern equatorial Africa but the case I would like to make is that it's part of us of a really a global signal that has to do with Pacific decadal oscillation variations so it's this what characterizes this persistence of dry now in Eastern equatorial Africa is part of this step change that happened somewhere at the end of the 1990s so this is an EOF that captures the broader picture but that also includes Eastern equatorial Africa similar so this is these are composites of sea surface temperature precipitation and winds that go with this shift so it's just simply the the difference in the averages of these two periods and again you see what you see what's popped out here is really the dominant I mean the one what seems to be the dominant influence here that can explain this shift is this Pacific signature or Indo-Pacific signature it's it can be I you know characterized as the first EOF in this residual sea surface temperature once you've removed Enso and global trend influences then whether you do this calculation on the Pacific or on the Indo-Pacific you get a time series like this red one here it's really the red and the blue so Pacific or Indo-Pacific give you the same multi-decadal variations here over the 20th century and so what we witnessed most recently is just the last shift here and so the shift I mean in some ways it's consistent with the El Nino influence in the sense that La Nina or cool conditions in the Eastern equatorial Pacific which occurred during La Nina are associated with dry Eastern equatorial Africa so I think that's all I have yeah so to wrap up there is a we can identify decadal variations in the oceans that have influence on these two regions and then we can also begin to tie two things together on the that allow us to go across time scales from a sub seasonal to multi-decadal